same, I've been liking https://openspec.dev/ and find the more time I put iterating/scrutinizing the spec artifacts (proposal, requirements, design, tasks) before I let the agent implement the better understanding I have and the better results I get. Also like that it is agent agnostic so I can take it with me as I try different agents/models.

Has there been a time when your spec was accurate, the agent implemented it, everything passed — but you still felt you didn't really understand what was happening? If so, what did you do about it? Did it degrade your ability to make subsequent specs?

sure, if I get lazy and don't read every line of the spec or don't thoroughly review the generated code, then I have felt a gap in understanding, but thats more on me than the tools. Another case is when I'm just completely unfamiliar with the framework the code is written in, I feel this happen whenever I use AI to write any non-trivial frontend as I rarely did frontend pre-AI.

Rated Analysis, rating relative chess analysis: https://rated-analysis.pickle-dev.com/

I've been a bit obsessed with chess lately and have read/heard things like "X is the top engine move, but no one under master plays that line". To me this begs the questions "which line are players below master playing" and "which line are players at/above/below my rating playing". This question motivated me to build a tool to analyze the move distribution (win/draw/loss rate) from a given position across rating buckets (400, 500, 600... 3000). Luckily, https://database.lichess.org/ provides an open/free database of games that have just the data required to build such a tool.

The Rated Analysis tool consists of an 1) an offline job to compute/store the positional move distribution bucketed by rating, and 2) the https://rated-analysis.pickle-dev.com/ webpage to serve the data along with some nice to haves like loading user games from platforms like chess.com/lichess.org.

Very interesting to see the above and below based on your own rating. A great add will be to show two columns: what players like me play here and the actual best counter after that move. An auto reply for the move i make will make this super useful and less cluttered than the usual tools.

thanks for taking a look!

> A great add will be to show two columns: what players like me play here and the actual best counter after that move

I think this case is handled by enabling the stockfish comparison toggle, but I may be misunderstanding.

> Claude Code uses Haiku to read through the transcript and decide if the goal has been completed.

feels kinda odd to use a less capable model to determine if the goal is fully complete. Especially if the user is expecting /goal to thoroughly complete the task. A less capable model would be more likely to misclassify `isComplete?`

> odd

There is nothing odd in the basic principle of Economics of using a "good enough" tool, optimal in other respects (e.g. cheap/er).

(You do not hire Nobel Prize winners as receptionists.)

The issue remains, whether the compromise to be adopted actually is "good enough".